man.c

   1 /*      $Id: man.c,v 1.123 2014/01/05 20:26:36 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010, 2011 Kristaps Dzonsons <kristaps@bsd.lv>
   4  * Copyright (c) 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
   5  *
   6  * Permission to use, copy, modify, and distribute this software for any
   7  * purpose with or without fee is hereby granted, provided that the above
   8  * copyright notice and this permission notice appear in all copies.
   9  *
  10  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  11  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  12  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  13  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  14  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  15  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  16  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  17  */
  18 #ifdef HAVE_CONFIG_H
  19 #include "config.h"
  20 #endif
  21
  22 #include <sys/types.h>
  23
  24 #include <assert.h>
  25 #include <stdarg.h>
  26 #include <stdlib.h>
  27 #include <stdio.h>
  28 #include <string.h>
  29
  30 #include "man.h"
  31 #include "mandoc.h"
  32 #include "libman.h"
  33 #include "libmandoc.h"
  34
  35 const   char *const __man_macronames[MAN_MAX] = {
  36         "br",           "TH",           "SH",           "SS",
  37         "TP",           "LP",           "PP",           "P",
  38         "IP",           "HP",           "SM",           "SB",
  39         "BI",           "IB",           "BR",           "RB",
  40         "R",            "B",            "I",            "IR",
  41         "RI",           "na",           "sp",           "nf",
  42         "fi",           "RE",           "RS",           "DT",
  43         "UC",           "PD",           "AT",           "in",
  44         "ft",           "OP",           "EX",           "EE",
  45         "UR",           "UE"
  46         };
  47
  48 const   char * const *man_macronames = __man_macronames;
  49
  50 static  struct man_node *man_node_alloc(struct man *, int, int,
  51                                 enum man_type, enum mant);
  52 static  int              man_node_append(struct man *,
  53                                 struct man_node *);
  54 static  void             man_node_free(struct man_node *);
  55 static  void             man_node_unlink(struct man *,
  56                                 struct man_node *);
  57 static  int              man_ptext(struct man *, int, char *, int);
  58 static  int              man_pmacro(struct man *, int, char *, int);
  59 static  void             man_free1(struct man *);
  60 static  void             man_alloc1(struct man *);
  61 static  int              man_descope(struct man *, int, int);
  62
  63
  64 const struct man_node *
  65 man_node(const struct man *man)
  66 {
  67
  68         assert( ! (MAN_HALT & man->flags));
  69         return(man->first);
  70 }
  71
  72
  73 const struct man_meta *
  74 man_meta(const struct man *man)
  75 {
  76
  77         assert( ! (MAN_HALT & man->flags));
  78         return(&man->meta);
  79 }
  80
  81
  82 void
  83 man_reset(struct man *man)
  84 {
  85
  86         man_free1(man);
  87         man_alloc1(man);
  88 }
  89
  90
  91 void
  92 man_free(struct man *man)
  93 {
  94
  95         man_free1(man);
  96         free(man);
  97 }
  98
  99
 100 struct man *
 101 man_alloc(struct roff *roff, struct mparse *parse, int quick)
 102 {
 103         struct man      *p;
 104
 105         p = mandoc_calloc(1, sizeof(struct man));
 106
 107         man_hash_init();
 108         p->parse = parse;
 109         p->quick = quick;
 110         p->roff = roff;
 111
 112         man_alloc1(p);
 113         return(p);
 114 }
 115
 116
 117 int
 118 man_endparse(struct man *man)
 119 {
 120
 121         assert( ! (MAN_HALT & man->flags));
 122         if (man_macroend(man))
 123                 return(1);
 124         man->flags |= MAN_HALT;
 125         return(0);
 126 }
 127
 128
 129 int
 130 man_parseln(struct man *man, int ln, char *buf, int offs)
 131 {
 132
 133         man->flags |= MAN_NEWLINE;
 134
 135         assert( ! (MAN_HALT & man->flags));
 136
 137         return (roff_getcontrol(man->roff, buf, &offs) ?
 138                         man_pmacro(man, ln, buf, offs) :
 139                         man_ptext(man, ln, buf, offs));
 140 }
 141
 142
 143 static void
 144 man_free1(struct man *man)
 145 {
 146
 147         if (man->first)
 148                 man_node_delete(man, man->first);
 149         if (man->meta.title)
 150                 free(man->meta.title);
 151         if (man->meta.source)
 152                 free(man->meta.source);
 153         if (man->meta.date)
 154                 free(man->meta.date);
 155         if (man->meta.vol)
 156                 free(man->meta.vol);
 157         if (man->meta.msec)
 158                 free(man->meta.msec);
 159 }
 160
 161
 162 static void
 163 man_alloc1(struct man *man)
 164 {
 165
 166         memset(&man->meta, 0, sizeof(struct man_meta));
 167         man->flags = 0;
 168         man->last = mandoc_calloc(1, sizeof(struct man_node));
 169         man->first = man->last;
 170         man->last->type = MAN_ROOT;
 171         man->last->tok = MAN_MAX;
 172         man->next = MAN_NEXT_CHILD;
 173 }
 174
 175
 176 static int
 177 man_node_append(struct man *man, struct man_node *p)
 178 {
 179
 180         assert(man->last);
 181         assert(man->first);
 182         assert(MAN_ROOT != p->type);
 183
 184         switch (man->next) {
 185         case (MAN_NEXT_SIBLING):
 186                 man->last->next = p;
 187                 p->prev = man->last;
 188                 p->parent = man->last->parent;
 189                 break;
 190         case (MAN_NEXT_CHILD):
 191                 man->last->child = p;
 192                 p->parent = man->last;
 193                 break;
 194         default:
 195                 abort();
 196                 /* NOTREACHED */
 197         }
 198
 199         assert(p->parent);
 200         p->parent->nchild++;
 201
 202         if ( ! man_valid_pre(man, p))
 203                 return(0);
 204
 205         switch (p->type) {
 206         case (MAN_HEAD):
 207                 assert(MAN_BLOCK == p->parent->type);
 208                 p->parent->head = p;
 209                 break;
 210         case (MAN_TAIL):
 211                 assert(MAN_BLOCK == p->parent->type);
 212                 p->parent->tail = p;
 213                 break;
 214         case (MAN_BODY):
 215                 assert(MAN_BLOCK == p->parent->type);
 216                 p->parent->body = p;
 217                 break;
 218         default:
 219                 break;
 220         }
 221
 222         man->last = p;
 223
 224         switch (p->type) {
 225         case (MAN_TBL):
 226                 /* FALLTHROUGH */
 227         case (MAN_TEXT):
 228                 if ( ! man_valid_post(man))
 229                         return(0);
 230                 break;
 231         default:
 232                 break;
 233         }
 234
 235         return(1);
 236 }
 237
 238
 239 static struct man_node *
 240 man_node_alloc(struct man *man, int line, int pos,
 241                 enum man_type type, enum mant tok)
 242 {
 243         struct man_node *p;
 244
 245         p = mandoc_calloc(1, sizeof(struct man_node));
 246         p->line = line;
 247         p->pos = pos;
 248         p->type = type;
 249         p->tok = tok;
 250
 251         if (MAN_NEWLINE & man->flags)
 252                 p->flags |= MAN_LINE;
 253         man->flags &= ~MAN_NEWLINE;
 254         return(p);
 255 }
 256
 257
 258 int
 259 man_elem_alloc(struct man *man, int line, int pos, enum mant tok)
 260 {
 261         struct man_node *p;
 262
 263         p = man_node_alloc(man, line, pos, MAN_ELEM, tok);
 264         if ( ! man_node_append(man, p))
 265                 return(0);
 266         man->next = MAN_NEXT_CHILD;
 267         return(1);
 268 }
 269
 270
 271 int
 272 man_tail_alloc(struct man *man, int line, int pos, enum mant tok)
 273 {
 274         struct man_node *p;
 275
 276         p = man_node_alloc(man, line, pos, MAN_TAIL, tok);
 277         if ( ! man_node_append(man, p))
 278                 return(0);
 279         man->next = MAN_NEXT_CHILD;
 280         return(1);
 281 }
 282
 283
 284 int
 285 man_head_alloc(struct man *man, int line, int pos, enum mant tok)
 286 {
 287         struct man_node *p;
 288
 289         p = man_node_alloc(man, line, pos, MAN_HEAD, tok);
 290         if ( ! man_node_append(man, p))
 291                 return(0);
 292         man->next = MAN_NEXT_CHILD;
 293         return(1);
 294 }
 295
 296
 297 int
 298 man_body_alloc(struct man *man, int line, int pos, enum mant tok)
 299 {
 300         struct man_node *p;
 301
 302         p = man_node_alloc(man, line, pos, MAN_BODY, tok);
 303         if ( ! man_node_append(man, p))
 304                 return(0);
 305         man->next = MAN_NEXT_CHILD;
 306         return(1);
 307 }
 308
 309
 310 int
 311 man_block_alloc(struct man *man, int line, int pos, enum mant tok)
 312 {
 313         struct man_node *p;
 314
 315         p = man_node_alloc(man, line, pos, MAN_BLOCK, tok);
 316         if ( ! man_node_append(man, p))
 317                 return(0);
 318         man->next = MAN_NEXT_CHILD;
 319         return(1);
 320 }
 321
 322 int
 323 man_word_alloc(struct man *man, int line, int pos, const char *word)
 324 {
 325         struct man_node *n;
 326
 327         n = man_node_alloc(man, line, pos, MAN_TEXT, MAN_MAX);
 328         n->string = roff_strdup(man->roff, word);
 329
 330         if ( ! man_node_append(man, n))
 331                 return(0);
 332
 333         man->next = MAN_NEXT_SIBLING;
 334         return(1);
 335 }
 336
 337
 338 /*
 339  * Free all of the resources held by a node.  This does NOT unlink a
 340  * node from its context; for that, see man_node_unlink().
 341  */
 342 static void
 343 man_node_free(struct man_node *p)
 344 {
 345
 346         if (p->string)
 347                 free(p->string);
 348         free(p);
 349 }
 350
 351
 352 void
 353 man_node_delete(struct man *man, struct man_node *p)
 354 {
 355
 356         while (p->child)
 357                 man_node_delete(man, p->child);
 358
 359         man_node_unlink(man, p);
 360         man_node_free(p);
 361 }
 362
 363 int
 364 man_addeqn(struct man *man, const struct eqn *ep)
 365 {
 366         struct man_node *n;
 367
 368         assert( ! (MAN_HALT & man->flags));
 369
 370         n = man_node_alloc(man, ep->ln, ep->pos, MAN_EQN, MAN_MAX);
 371         n->eqn = ep;
 372
 373         if ( ! man_node_append(man, n))
 374                 return(0);
 375
 376         man->next = MAN_NEXT_SIBLING;
 377         return(man_descope(man, ep->ln, ep->pos));
 378 }
 379
 380 int
 381 man_addspan(struct man *man, const struct tbl_span *sp)
 382 {
 383         struct man_node *n;
 384
 385         assert( ! (MAN_HALT & man->flags));
 386
 387         n = man_node_alloc(man, sp->line, 0, MAN_TBL, MAN_MAX);
 388         n->span = sp;
 389
 390         if ( ! man_node_append(man, n))
 391                 return(0);
 392
 393         man->next = MAN_NEXT_SIBLING;
 394         return(man_descope(man, sp->line, 0));
 395 }
 396
 397 static int
 398 man_descope(struct man *man, int line, int offs)
 399 {
 400         /*
 401          * Co-ordinate what happens with having a next-line scope open:
 402          * first close out the element scope (if applicable), then close
 403          * out the block scope (also if applicable).
 404          */
 405
 406         if (MAN_ELINE & man->flags) {
 407                 man->flags &= ~MAN_ELINE;
 408                 if ( ! man_unscope(man, man->last->parent, MANDOCERR_MAX))
 409                         return(0);
 410         }
 411
 412         if ( ! (MAN_BLINE & man->flags))
 413                 return(1);
 414         man->flags &= ~MAN_BLINE;
 415
 416         if ( ! man_unscope(man, man->last->parent, MANDOCERR_MAX))
 417                 return(0);
 418         return(man_body_alloc(man, line, offs, man->last->tok));
 419 }
 420
 421 static int
 422 man_ptext(struct man *man, int line, char *buf, int offs)
 423 {
 424         int              i;
 425
 426         /* Literal free-form text whitespace is preserved. */
 427
 428         if (MAN_LITERAL & man->flags) {
 429                 if ( ! man_word_alloc(man, line, offs, buf + offs))
 430                         return(0);
 431                 return(man_descope(man, line, offs));
 432         }
 433
 434         for (i = offs; ' ' == buf[i]; i++)
 435                 /* Skip leading whitespace. */ ;
 436
 437         /*
 438          * Blank lines are ignored right after headings
 439          * but add a single vertical space elsewhere.
 440          */
 441
 442         if ('\0' == buf[i]) {
 443                 /* Allocate a blank entry. */
 444                 if (MAN_SH != man->last->tok &&
 445                     MAN_SS != man->last->tok) {
 446                         if ( ! man_elem_alloc(man, line, offs, MAN_sp))
 447                                 return(0);
 448                         man->next = MAN_NEXT_SIBLING;
 449                 }
 450                 return(1);
 451         }
 452
 453         /*
 454          * Warn if the last un-escaped character is whitespace. Then
 455          * strip away the remaining spaces (tabs stay!).
 456          */
 457
 458         i = (int)strlen(buf);
 459         assert(i);
 460
 461         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 462                 if (i > 1 && '\\' != buf[i - 2])
 463                         man_pmsg(man, line, i - 1, MANDOCERR_EOLNSPACE);
 464
 465                 for (--i; i && ' ' == buf[i]; i--)
 466                         /* Spin back to non-space. */ ;
 467
 468                 /* Jump ahead of escaped whitespace. */
 469                 i += '\\' == buf[i] ? 2 : 1;
 470
 471                 buf[i] = '\0';
 472         }
 473
 474         if ( ! man_word_alloc(man, line, offs, buf + offs))
 475                 return(0);
 476
 477         /*
 478          * End-of-sentence check.  If the last character is an unescaped
 479          * EOS character, then flag the node as being the end of a
 480          * sentence.  The front-end will know how to interpret this.
 481          */
 482
 483         assert(i);
 484         if (mandoc_eos(buf, (size_t)i))
 485                 man->last->flags |= MAN_EOS;
 486
 487         return(man_descope(man, line, offs));
 488 }
 489
 490 static int
 491 man_pmacro(struct man *man, int ln, char *buf, int offs)
 492 {
 493         int              i, ppos;
 494         enum mant        tok;
 495         char             mac[5];
 496         struct man_node *n;
 497
 498         if ('"' == buf[offs]) {
 499                 man_pmsg(man, ln, offs, MANDOCERR_BADCOMMENT);
 500                 return(1);
 501         } else if ('\0' == buf[offs])
 502                 return(1);
 503
 504         ppos = offs;
 505
 506         /*
 507          * Copy the first word into a nil-terminated buffer.
 508          * Stop copying when a tab, space, or eoln is encountered.
 509          */
 510
 511         i = 0;
 512         while (i < 4 && '\0' != buf[offs] &&
 513                         ' ' != buf[offs] && '\t' != buf[offs])
 514                 mac[i++] = buf[offs++];
 515
 516         mac[i] = '\0';
 517
 518         tok = (i > 0 && i < 4) ? man_hash_find(mac) : MAN_MAX;
 519
 520         if (MAN_MAX == tok) {
 521                 mandoc_vmsg(MANDOCERR_MACRO, man->parse, ln,
 522                                 ppos, "%s", buf + ppos - 1);
 523                 return(1);
 524         }
 525
 526         /* The macro is sane.  Jump to the next word. */
 527
 528         while (buf[offs] && ' ' == buf[offs])
 529                 offs++;
 530
 531         /*
 532          * Trailing whitespace.  Note that tabs are allowed to be passed
 533          * into the parser as "text", so we only warn about spaces here.
 534          */
 535
 536         if ('\0' == buf[offs] && ' ' == buf[offs - 1])
 537                 man_pmsg(man, ln, offs - 1, MANDOCERR_EOLNSPACE);
 538
 539         /*
 540          * Remove prior ELINE macro, as it's being clobbered by a new
 541          * macro.  Note that NSCOPED macros do not close out ELINE
 542          * macros---they don't print text---so we let those slip by.
 543          */
 544
 545         if ( ! (MAN_NSCOPED & man_macros[tok].flags) &&
 546                         man->flags & MAN_ELINE) {
 547                 n = man->last;
 548                 assert(MAN_TEXT != n->type);
 549
 550                 /* Remove repeated NSCOPED macros causing ELINE. */
 551
 552                 if (MAN_NSCOPED & man_macros[n->tok].flags)
 553                         n = n->parent;
 554
 555                 mandoc_vmsg(MANDOCERR_LINESCOPE, man->parse, n->line,
 556                     n->pos, "%s breaks %s", man_macronames[tok],
 557                     man_macronames[n->tok]);
 558
 559                 man_node_delete(man, n);
 560                 man->flags &= ~MAN_ELINE;
 561         }
 562
 563         /*
 564          * Remove prior BLINE macro that is being clobbered.
 565          */
 566         if ((man->flags & MAN_BLINE) &&
 567             (MAN_BSCOPE & man_macros[tok].flags)) {
 568                 n = man->last;
 569
 570                 /* Might be a text node like 8 in
 571                  * .TP 8
 572                  * .SH foo
 573                  */
 574                 if (MAN_TEXT == n->type)
 575                         n = n->parent;
 576
 577                 /* Remove element that didn't end BLINE, if any. */
 578                 if ( ! (MAN_BSCOPE & man_macros[n->tok].flags))
 579                         n = n->parent;
 580
 581                 assert(MAN_HEAD == n->type);
 582                 n = n->parent;
 583                 assert(MAN_BLOCK == n->type);
 584                 assert(MAN_SCOPED & man_macros[n->tok].flags);
 585
 586                 mandoc_vmsg(MANDOCERR_LINESCOPE, man->parse, n->line,
 587                     n->pos, "%s breaks %s", man_macronames[tok],
 588                     man_macronames[n->tok]);
 589
 590                 man_node_delete(man, n);
 591                 man->flags &= ~MAN_BLINE;
 592         }
 593
 594         /*
 595          * Save the fact that we're in the next-line for a block.  In
 596          * this way, embedded roff instructions can "remember" state
 597          * when they exit.
 598          */
 599
 600         if (MAN_BLINE & man->flags)
 601                 man->flags |= MAN_BPLINE;
 602
 603         /* Call to handler... */
 604
 605         assert(man_macros[tok].fp);
 606         if ( ! (*man_macros[tok].fp)(man, tok, ln, ppos, &offs, buf))
 607                 goto err;
 608
 609         /* In quick mode (for mandocdb), abort after the NAME section. */
 610
 611         if (man->quick && MAN_SH == tok &&
 612             strcmp(man->last->prev->child->string, "NAME"))
 613                 return(2);
 614
 615         /*
 616          * We weren't in a block-line scope when entering the
 617          * above-parsed macro, so return.
 618          */
 619
 620         if ( ! (MAN_BPLINE & man->flags)) {
 621                 man->flags &= ~MAN_ILINE;
 622                 return(1);
 623         }
 624         man->flags &= ~MAN_BPLINE;
 625
 626         /*
 627          * If we're in a block scope, then allow this macro to slip by
 628          * without closing scope around it.
 629          */
 630
 631         if (MAN_ILINE & man->flags) {
 632                 man->flags &= ~MAN_ILINE;
 633                 return(1);
 634         }
 635
 636         /*
 637          * If we've opened a new next-line element scope, then return
 638          * now, as the next line will close out the block scope.
 639          */
 640
 641         if (MAN_ELINE & man->flags)
 642                 return(1);
 643
 644         /* Close out the block scope opened in the prior line.  */
 645
 646         assert(MAN_BLINE & man->flags);
 647         man->flags &= ~MAN_BLINE;
 648
 649         if ( ! man_unscope(man, man->last->parent, MANDOCERR_MAX))
 650                 return(0);
 651         return(man_body_alloc(man, ln, ppos, man->last->tok));
 652
 653 err:    /* Error out. */
 654
 655         man->flags |= MAN_HALT;
 656         return(0);
 657 }
 658
 659 /*
 660  * Unlink a node from its context.  If "man" is provided, the last parse
 661  * point will also be adjusted accordingly.
 662  */
 663 static void
 664 man_node_unlink(struct man *man, struct man_node *n)
 665 {
 666
 667         /* Adjust siblings. */
 668
 669         if (n->prev)
 670                 n->prev->next = n->next;
 671         if (n->next)
 672                 n->next->prev = n->prev;
 673
 674         /* Adjust parent. */
 675
 676         if (n->parent) {
 677                 n->parent->nchild--;
 678                 if (n->parent->child == n)
 679                         n->parent->child = n->prev ? n->prev : n->next;
 680         }
 681
 682         /* Adjust parse point, if applicable. */
 683
 684         if (man && man->last == n) {
 685                 /*XXX: this can occur when bailing from validation. */
 686                 /*assert(NULL == n->next);*/
 687                 if (n->prev) {
 688                         man->last = n->prev;
 689                         man->next = MAN_NEXT_SIBLING;
 690                 } else {
 691                         man->last = n->parent;
 692                         man->next = MAN_NEXT_CHILD;
 693                 }
 694         }
 695
 696         if (man && man->first == n)
 697                 man->first = NULL;
 698 }
 699
 700 const struct mparse *
 701 man_mparse(const struct man *man)
 702 {
 703
 704         assert(man && man->parse);
 705         return(man->parse);
 706 }